KBrief flashback: From April 15 to October 31, 1980, around 150,000 Cubans left their homeland, half of them settled in Miami. Most of them got residence and work permits. As a result, the number of workers in the big city grew by 7 percent in a very short time. The local unemployment rate rose from 5 to 7 percent, but quickly normalized.
The economics professor David Card, who has now been awarded the Nobel Memorial Prize for Economics, examined wage developments in Miami from 1979 to 1985 and compared them with other American metropolises. Simple economic models suggested that wages would generally decline, and particularly sharply for workers who competed directly with Cubans. Card, on the other hand, found that immigration had little impact on wages.
The award from the Norwegian Prize Committee honors less the surprising result than the method that is comparatively new for economics: Card used a natural special event to reach conclusions. The Berkeley researcher, who received half the award, shares it with Guido Imbens (Stanford University) and Joshua Angrist (Massachusetts Institute of Technology). What all three researchers have in common is that they use so-called natural experiments to test theories. Together with the late Alan Krueger (Princeton), Angrist exploited a peculiarity in the American school system to show that more schools produce more income, while statistics specialist Imbens is primarily characterized by new methodological findings.
Machine learning in economics
Card’s second groundbreaking work, in turn, changed ideas about the minimum wage. Also with Krueger, he investigated the effect that the rise in the statutory minimum wage in the state of New Jersey had on employment in the fast-food restaurants compared to neighboring Pennsylvania, where the minimum wages were lower and not adjusted. Contrary to the “classical” theory, they found no negative employment effect; rather, employment increased in the more than 400 fast-service restaurants in New Jersey that were examined. This was also surprising because the wage increase took place during a recession.
Card’s minimum wage research is currently being tried mainly by sections of the Democratic Party to support the demand for a minimum hourly wage of $ 15. Card found that this work was often misunderstood and misused by left and right. He himself had never called for a general increase in the minimum wage. In the particular case he was investigating, New Jersey raised the minimum wage in 1992 from $ 4.25 to $ 5.05, while in Pennsylvania it remained stable at $ 4.25. In 1992, $ 5 had purchasing power that is less than $ 10 today. Card, as he once stated, quickly left the field of research not only because he lost good friends at the University of Chicago, for example, but also because he did not want to dwell on commenting on old results for a researcher’s life. Instead, when interpreting data from Germany, he found out that women also earn less than men because they are more often employed by employers that generally pay less.
While Card was primarily convincing through its more detailed analysis of the labor market, the Nobel Prize Committee chose Angrist and Imbens for how they further developed the theoretical foundations of statistical analysis. Their contributions, published in the 1990s, also opened up new possibilities for economists to draw conclusions from observed data. In fact, a large part of current statistical methods is concerned with uncovering connections between different phenomena in the first place, i.e. with clarifying whether and how they are correlated. The question that is just as important in practice for economic politicians or companies is how the observable relationship runs, what is the cause and what is the effect, whether and how a result actually changes when the framework conditions change or whether other variables play a role, which are not directly mapped at all. It’s about causality as opposed to mere correlation. Here Angrist and Imbens have provided new techniques that researchers around the world apply in their empirical work in a variety of ways.
Imbens has also made a name for itself with his urge to transfer data-intensive approaches based on learning algorithms from artificial intelligence into economics. Together with his wife Susan Athey, who is occasionally also mentioned as a Nobel Prize-worthy researcher, he published a successful overview article two years ago under the title “Machine Learning Methods Economists Should Know About”. The drive is clear: At a time when huge amounts of data are constantly being collected and evaluated (“big data”), economics is increasingly being linked in parts with computer science. It is not surprising that the name Judea Pearl appears in the literature appendix to the explanation of the award committee: In 2011, the computer scientist received the highest distinction in his subject, the Turing Award – for his work on causality.